Implement a helper function for rebasing base-url (to be used with cli)

MarkBind · May 16, 2017 · 64b2ff6 · 64b2ff6
1 parent 4017689
commit 64b2ff6
Show file tree

Hide file tree

Showing 3 changed files with 915 additions and 13 deletions.
diff --git a/lib/parser.js b/lib/parser.js
@@ -15,12 +15,15 @@ const utils = require('./utils');
 const fs = require('fs');
 const path = require('path');
 const url = require('url');
+const nunjucks = require('nunjucks');
 
+const ATTRIB_INCLUDE_PATH = 'include-path';
+const ATTRIB_CWF = 'cwf';
 /*
  * Utils
  */
 function isText(element) {
-  return element.type === 'text';
+  return element.type === 'text' || element.type === 'comment';
 }
 
 function Parser(options) {
@@ -36,6 +39,8 @@ Parser.prototype.getDynamicIncludeSrc = function () {
 
 Parser.prototype._preprocess = function (element, context) {
   let self = this;
+  element.attribs = element.attribs || {};
+  element.attribs[ATTRIB_CWF] = path.resolve(context.cwf);
   if (element.name === 'include') {
     let isInline = _.hasIn(element.attribs, 'inline');
     let isDynamic = _.hasIn(element.attribs, 'dynamic');
@@ -44,6 +49,7 @@ Parser.prototype._preprocess = function (element, context) {
     let includeSrcPath = includeSrc.path;
     let filePath = isUrl ? element.attribs.src : path.resolve(path.dirname(context.cwf), includeSrcPath);
     element.name = isInline ? 'span' : 'div';
+    element.attribs[ATTRIB_INCLUDE_PATH] = filePath;
 
     if (isDynamic) {
       element.name = 'dynamic-panel';
@@ -370,4 +376,117 @@ Parser.prototype.renderFile = function (file, cb) {
   });
 };
 
+Parser.prototype.resolveBaseUrl = function (pageData, config, cb) {
+  let {baseUrlMap, rootPath} = config;
+  this.baseUrlMap = baseUrlMap;
+  this.rootPath = rootPath;
+  cb = cb || function () {}; // create empty callback
+  return new Promise((resolve, reject) => {
+    let handler = new htmlparser.DomHandler((error, dom) => {
+      if (error) {
+        reject(error);
+        cb(error);
+        return;
+      }
+      dom.forEach(d => {
+        let childrenBase = {};
+        this._rebaseReference(d, childrenBase);
+        if (d.attribs && d.attribs[ATTRIB_CWF] && d.attribs[ATTRIB_INCLUDE_PATH]) {
+          let currentBase = calculateNewBaseUrl(d.attribs[ATTRIB_CWF], this.rootPath, this.baseUrlMap);
+          let newBase = calculateNewBaseUrl(d.attribs[ATTRIB_CWF], this.rootPath, this.baseUrlMap);
+          if (currentBase.relative !== newBase.relative) {
+            let bases = Object.keys(childrenBase);
+            if (bases.length !== 0) {
+              // need to rebase
+              let children = d.children;
+              if (children) {
+                cheerio.prototype.options.xmlMode = false;
+                let rendered = nunjucks.renderString(cheerio.html(children), {baseUrl: `{{baseUrl}}/${newBase.relative}`});
+                d.children = cheerio.parseHTML(rendered, true);
+                cheerio.prototype.options.xmlMode = true;
+              }
+            }
+          }
+        }
+      });
+      cheerio.prototype.options.xmlMode = false;
+      resolve(cheerio.html(dom));
+      cb(null, cheerio.html(dom));
+      cheerio.prototype.options.xmlMode = true;
+    });
+
+    let parser = new htmlparser.Parser(handler, {
+      xmlMode: true,
+      decodeEntities: true
+    });
+
+    parser.parseComplete(pageData);
+  });
+};
+
+Parser.prototype._rebaseReference = function(node, foundBase) {
+  if (_.isArray(node)) {
+    return node.forEach(el => {
+      return this._rebaseReference(el, foundBase);
+    });
+  }
+
+  if (isText(node)) {
+    return node;
+  } else {
+    // Rebase children element
+    let childrenBase = {};
+    node.children.forEach(el => {
+      this._rebaseReference(el, childrenBase);
+    });
+
+    // rebase current element
+    if (node.attribs[ATTRIB_CWF]) {
+      let path = node.attribs[ATTRIB_CWF];
+      let newBase = calculateNewBaseUrl(path, this.rootPath, this.baseUrlMap);
+      if (newBase) {
+        let {relative, parent} = newBase;
+        foundBase[parent] = relative;
+      }
+
+      let bases = Object.keys(childrenBase);
+      if (bases.length !== 0) {
+        // need to rebase
+        let newBase = childrenBase[bases[0]];
+        let children = node.children;
+        if (children) {
+          let currentBase = calculateNewBaseUrl(node.attribs[ATTRIB_CWF], this.rootPath, this.baseUrlMap);
+          if (currentBase) {
+            if (currentBase.relative !== newBase) {
+              cheerio.prototype.options.xmlMode = false;
+              let rendered = nunjucks.renderString(cheerio.html(children), {baseUrl: `{{baseUrl}}/${newBase}`});
+              node.children = cheerio.parseHTML(rendered, true);
+              cheerio.prototype.options.xmlMode = true;
+            }
+          }
+        }
+      }
+      delete node.attribs[ATTRIB_INCLUDE_PATH];
+    }
+    delete node.attribs[ATTRIB_CWF];
+  }
+};
+
+function calculateNewBaseUrl(filePath, root, lookUp) {
+  function calculate(file, result) {
+    if (file === root) {
+      return {relative: path.relative(root, root), parent: root};
+    }
+    let parent = path.dirname(file);
+    if (lookUp[parent] && result.length == 1) {
+      return {relative: path.relative(parent, result[0]), parent};
+    } else if (lookUp[parent]) {
+      return calculate(parent, [parent]);
+    }
+    return calculate(parent, result);
+  }
+
+  return calculate(filePath, []);
+}
+
 module.exports = Parser;
diff --git a/package.json b/package.json
@@ -26,7 +26,9 @@
     "markdown-it-mark": "^2.0.0",
     "markdown-it-table-of-contents": "^0.3.2",
     "markdown-it-task-lists": "^1.4.1",
-    "markdown-it-video": "^0.4.0"
+    "markdown-it-video": "^0.4.0",
+    "nunjucks": "^3.0.0",
+    "path-is-inside": "^1.0.2"
   },
   "devDependencies": {
     "eslint": "^3.16.1",